Create Your First Deployment

TrueFoundry helps you to seamlessly manage the entire machine learning lifecycle, from experimentation to deployment and beyond. You can:

1. Kickstart your machine learning journey by launching a Jupyter Notebook to explore and experiment with your ideas.

   • Launch a Notebook
   • Launch SSH Server and Connect with VS Code

2. Once your model is ready for training, execute a model training job from within the Notebook using the Python SDK. Or you can push your training code to a Github Repository and deploy directly from a public Github repository

   • Deploy a Job from Github Repo
   • Deploy a Job using Python SDK

3. Seamlessly log your trained model to the TrueFoundry Model Registry, which is backed by a secure blob storage service like S3, GCS, or Azure Container.

   • Log your trained model

4. Deploy the logged model as a:

   1. Real-time API Service: Deploy your model as a real-time API Service to serve predictions in real-time, either from a public Github repository or from a local-machine / notebook

      • Deploy a Service from Github Repo
      • Deploy a Service using Python SDK

   2. Batch Inference Service: Deploy your model for batch inference to process large datasets efficiently by deploying it as a Job

   3. Async Service: Handle requests asynchronously using a queue to store intermediate requests by deploying an Async Service

      • Deploy an Async Service

5. LLM Testing and Deployment: Evaluate and compare the performance of various LLMs using TrueFoundry’s AI Gateway capabilities. Once you’ve selected the desired LLM, deploy it with ease using pre-configured settings

   • Compare various LLM’s using AI Gateway
   • Deploy a LLM

6. LLM Finetuning: Leverage TrueFoundry’s LLM finetuning capabilities to tailor LLMs to your specific needs and data.

   • Finetune a LLM on your own data